01
Univariate analysis: centre, spread, shape, outliers
Mean ten lakh, median six. Somebody in that dataset earns a great deal more than everyone else.
One column at a time: central values (mean, median, mode), spread (standard deviation, quartiles), shape and outliers. A mean far above the median signals high outliers. Charts: box plot, histogram (a bar chart grouped into bins), density plot; skewness (ages mostly 25-40 with a few over 70 is right-skewed). On the orders dataset: counts by status and platform, most frequent category, size and city, maximum and average amount. This is the baseline you prompt an AI tool for first.
Why it matters
Every later prompt starts with 'univariate, bivariate, multivariate'.
This idea elsewhereparallelKairos terminology lesson